Stroke is a debilitating neurological disorder that is associated with tremendous socioeconomic burden. Whereas FDA-approved tissue plasminogen activator (tPA) is a potent thrombolytic agent for treating acute ischemic stroke, few patients present for treatment within its therapeutic window. There remains substantial interest in developing strategies to extend the thrombolytic therapy into subpopulations who may benefit from late intervention. MRI has become an essential clinical tool for the triage and management of acute ischemic stroke patients, yet the conventional stroke MRI is inadequate to fully characterize the heterogeneous ischemic tissue injury and effectively guide treatment in late presenting stroke patients. The approximation of diffusion/perfusion MRI (DWI/PWI) mismatch as ischemic penumbra, despite its initial enthusiasm, has been recognized to be oversimplified. PWI lesion contains tissue at no risk to infarction, while DWI lesion may recover if promptly reperfused. Besides falsely characterizing some salvageable tissue as ischemic core, DWI may also fail to identify tissue that has already suffered irreversible injury, as the eventual infarction is often larger than acute DWI lesion. As noted in the report of NIH/NINDS Stroke Progress Review Group (SPRG) in 2011, the number one priority for stroke imaging is to understand the impact of hemodynamics, collateral flow, oxygen and brain metabolism upon tissue survival and function. Tissue acidosis is closely associated with tissue oxygen/glucose metabolism, and may provide a metabolic biomarker for defining ischemic penumbra. However, currently available in vivo pH measurement techniques have significant limitations. Our proposal aims to develop endogenous amide proton chemical exchange saturation transfer (CEST) MRI for fast and non-invasive pH imaging. We will first develop novel acquisition and post-processing strategies to enhance the sensitivity of CEST imaging (Aim1). We will develop quantitative analysis that transform pH-weighted MRI to absolute tissue pH mapping in experimental stroke model, and test it under varied glycemic conditions and stroke onset time (Aim 2). We will then evaluate pH imaging, a novel metabolic imaging marker, to guide tPA thrombolysis in an embolic stroke model that mimics tPA thrombolysis in patients (Aim 3). In summary, our proposal establishes fast and quantitative pH stroke imaging in experimental stroke models, and once the sensitivity and specificity of pH MRI in defining metabolic penumbra are confirmed, we will translate it to clinic and evaluate its utility in late-presenting stroke patients.